The answer was on the next bill

The tariff code was only four digits: 7200. For a food processor in Landsborough, those four digits helped determine what it paid to get electricity through the network. The question was whether they still described the best arrangement for the way the business actually used power.

We started with the bills and a year's worth of interval data. The pattern mattered: how much electricity the site used, when it used it and how high its demand went. We compared that pattern with the available network tariffs. The review supported a move from Energex tariff 7200 to 8300, and the change went ahead.

That was encouraging. But a recommendation on a spreadsheet still had to turn into a difference on an invoice. So we waited for the bill, then checked it.

July's network charges came to $13,237. For the same month's consumption and demand, we calculated that the old tariff would have cost $14,101. That was $864 less, excluding GST. The calculation for the new tariff closely matched what the retailer had actually billed.

Then we checked August. The network charges were $13,191, compared with $14,009 modelled on the old tariff. Allowing for rounding, that was another $819 less. Across the two months, the difference was $1,683, or about 6% of the network charges.

Those figures cover two months of network charges, not a full year or the whole electricity bill. They were enough to answer the question we had started with. The tariff had changed, the bills reflected it, and the business was paying less for the same electricity use. Four digits were worth a second look.